[atlas]P466 - inconsistent readings?
Robert Kisteleki robert at ripe.net
Tue Nov 30 16:12:32 CET 2010
On 2010.11.30. 13:40, Wilfried Woeber, UniVie/ACOnet wrote: >> Your graphs look fine to me now. Maybe I missed the window of opportunity? :-) > > Maybe :-) > > My point is that the "outage panel" lists a lot of "downs" for yesterday, > but these periods are not visible in the graphs for that timeframe as grey > "no data" bars. Given the stability of my ADSL link, I'd rather believe in > the graphs instead of the list of down-up transitions. > > So, I probably should have asked: which criteria do you select or evaluate > to declare a probe "down" and then "up" again? :-) A probe is considered to be "up" if it's connected to our infrastructure and is sending in measurement data. If it isn't, then it's "down". If the probe disconnects for some reason, it'll try to get back as soon as it can. This is recorded as an event in the connection log. However, if the probe is able to get back quickly and resume measurements, there may not be gaps at all in the graphs. So, observing some downtime and having continuous graphs are not mutually exclusive. > Btw, just to make that clear, this is not a complaint or an attempt to bother > you excessively. I "simply" try to understand how this stuff works. > > Because I consider this understanding as essential to grok the information in > the graphs (and to be able to make potential probe-sponsors intersted). Therefore your questions are perfectly fair, and I hope I or other members of the team can answer them sufficiently. Cheers, Robert > Cheers, > Wilfried.